1. Field of the Invention
The present invention relates to image sensors, and in particular, to image sensors with pixel circuits implemented with metal-insulator-semiconductor (MIS) photodiodes.
2. Description of the Related Art
Image sensors, such as those used for large area X-ray imaging, often use pixel circuits in which a mesa-isolated MIS photodiode is used as the photosensitive device. (A mesa-isolated device is formed by etching away a portion of the active materials, leaving a “mesa” of active materials.) Another common photosensitive device is a mesa-isolated p-i-n photodiode. Yet another conventional photosensitive device is a p-i-n photodiode comprised substantially of continuous films. However, such conventional photosensitive devices have disadvantages. Both the mesa-isolated MIS and p-i-n photodiodes typically generate poor image signals. The p-i-n photodiode comprised substantially of continuous films exhibits significant crosstalk among adjacent pixels.
Referring to FIG. 1, a conventional embodiment 10a of a pixel circuit implemented with a mesa-isolated MIS photodiode and a thin-film transistor (TFT) is typically integrated as shown. Starting with a substrate 12, various layers of dielectric (insulator), semiconductor and conductive materials are formed (e.g., deposited). For example, on the top surface of the substrate 12, a patterned layer of conductive material (e.g., metal) forms the bottom electrode 20a of the MIS photodiode 14a and the gate terminal 32 of the TFT 16. Next is a patterned layer of dielectric material which forms the dielectric 26a of the MIS photodiode 14a and the gate dielectric 34 of the TFT 16. Next is a patterned layer of intrinsic amorphous silicon (i a-Si) material which forms one of the semiconductor layers 24a, the light absorbing layer, of the MIS photodiode 14a and the channel 36 of the TFT 16. Next is a patterned layer of n+ amorphous silicon which forms the remaining semiconductor layer, the ohmic contact 22a, and, effectively, the top electrode of the MIS photodiode 14a and ohmic contacts 38 for the drain and source terminals of the TFT 16. Next is another patterned conductive layer (e.g., metal) which forms the drain 42 and source 44 terminals of the TFT 16, the data line 46, and the bias line 30. Following all of this is a layer of passivation (dielectric) 50.
Referring to FIG. 2, another embodiment 10b of a conventional pixel circuit uses a mesa-isolated p-i-n photodiode 14b instead of a mesa-isolated MIS photodiode. In this embodiment 10b, the structure of the TFT 16 is substantially the same as the embodiment 10a of FIG. 1. In place of the MIS photodiode 14a, however, a p-i-n photodiode 14b is used. The patterned layer of conductive material (e.g., metal) forming the source terminal 44 of the TFT 16 also forms the bottom electrode 20b of the p-i-n photodiode 14b. Next is a patterned layer of n+ amorphous silicon 28b, followed by a patterned layer of intrinsic amorphous silicon 24b, the light absorbing layer, and then a patterned layer of p+ amorphous silicon 22b which together form the p-i-n structure of the photodiode 14b. Next is a patterned layer of optically transparent conductive material (e.g., indium tin oxide, or ITO) which forms the top electrode 18b. Next is a patterned layer of dielectric material which forms an interlayer dielectric 52, through which a via is formed to allow conductive material (e.g., metal) to be deposited to form the bias line 30 in contact with the top electrode 18b of the photodiode 14b. Lastly is a layer of passivation 50.
Referring to FIG. 3, an alternative embodiment 10c of a conventional pixel circuit using a p-i-n photodiode 14c is similar to the embodiment 10b of FIG. 2, except that a substantial portion of the photodiode 14c is formed by using continuous films, as opposed to being formed in a mesa-isolated structure. Accordingly, the fabrication and materials used for the various photodiode layers 24c, 22c and 18c are the same, but in a continuous film.
As noted above, a disadvantage common to mesa-isolated MIS and p-i-n photodiode sensors is low signal levels. With a mesa-isolated structure, such photosensitive elements have fill-factors less than unity (fill-factor is defined as the area of the photosensitive element divided by the overall pixel area). Hence, not all of the light impinging upon the pixel is absorbed by the photosensitive element. Accordingly, maximum possible signal strength cannot be achieved.
The mesa-isolated MIS photodiode structure of FIG. 1 has a further disadvantage. The same film that is used to form the channel 36 of the TFT 16 is also used to form the light absorbing layer 24a of the MIS photodiode 14a. Generally, TFT 16 performance is optimized when the channel 36 thickness is thin, while MIS photodiode 14a performance is optimized when the light absorbing layer 24a is thick. With a single film, the performance of one or both of the TFT 16 and MIS photodiode 14a may suffer as the chosen film thickness may not be optimum for one or both.
With respect to signal strength, the p-i-n photodiode 14c formed substantially of continuous films, as shown in the embodiment 10c of FIG. 3, has improved signal strength. With this photosensitive element having a near unity fill-factor, nearly maximum signal strength can be achieved. However, this structure can suffer from significant crosstalk between adjacent pixels. For example, the interface 54 between the interlayer dielectric 52 and the light absorption layer 24c can have a nonzero conductance. Accordingly, potential differences between the bottom electrodes 20c of adjacent pixels produce small currents between such pixels, i.e., crosstalk.